altllm-portal-billing

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.65). The required workflow for the credit/history/usage commands reads Portal API responses (including arbitrary JSON/body text from the platform at runtime) and passes them through to the CLI/LLM context, which is effectively outsider-authored free text because it originates from an external service the user didn’t author.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill exposes Portal billing commands (e.g., redeem-promo) that perform authenticated billing actions using the saved Portal session token, allowing modification of account credit/balance. This is a purpose-built billing/finance interface (not a generic tool), so it grants direct financial execution capability.